Why do voters forgive corrupt politicians? Cynicism, noise and implicit exchange

Why do voters forgive corrupt politicians? Cynicism, noise and implicit exchange Jordi Muñoz, Eva Anduiza and Aina Gallego Universitat Autònoma de Barcelona Jordi.Munoz@uab.cat Eva.Anduiza@uab.cat Aina.Gallego@uab.cat Paper prepared to be presented at the IPSA conference, Madrid, July 2012 Very preliminary draft, comments welcome. 1

Abstract Contrarily to what democratic theory would expect and what voters tend to declare in surveys, evidence shows limited electoral consequences for corrupt misbehaviour of politicians. Why do voters forgive corrupt politicians when it comes to voting? Previous research has highlighted the conditioning role of (lack of) information, (weak) institutions and partisanship in explaining this fact. In this paper we propose three micro mechanisms that could explain the fact that corruption has limited electoral consequences even in contexts with high information and strong institutions: implicit exchange (voters take into account other elements of the politicians or parties performance that counterbalance the effects of corruption), noise (voters perceive corruption allegations as partisan tricks with limited credibility), and cynicism (voters think that any politician would behave in a corrupt way). We test these three mechanisms with a survey experiment. Our findings show that both implicit exchange and noise are important factors explaining voters support for corrupt politicians. We fail to find an effect of cynicism. 2

Keywords: corruption, voting, experiment Introduction Researchers have often been puzzled about the limited effects of corruption on voting and election results. It is the widely observed paradox: unpopular corruption and popular corrupt politicians (Kurer 2001, pp. 63). This paradox, that puts democratic accountability into question, has been observed in different countries and contexts: the US (Rundquist et al, 1977; Dimock and Jacobson, 1995), the UK (Eggers & Fisher, 2011), Japan (Reed, 1999), Italy (Chang, Golden and Hill, 2010), Spain (Rivero- Rodriguez and Fernandez-Vazquez, 2011). The puzzle is so startling for democratic theory that some scholars have defined large sets of conditions under which corruption could be expected to have electoral consequences (Jiménez and Caínzos, 2004). Within this perspective, previous research has found that information, institutions and partisanship play a significant role in explaining why people do not systematically throw corrupt politicians out. Information seems to be a necessary, though not a sufficient condition for citizens to punish corrupt politicians (Figueiredo et al., 2000, Cong, De la O, Karlan & Wantchekon, 2011). Partisans are less likely to perceive and punish corruption among their politicians and public officers (Dimock & Jacobson, 1995; Gonzales et al 1995; Anduiza, Gallego & Muñoz, 2014). And some institutions may further reduce the electoral consequences of corruption (Manzetti & Wilson 2007). However, we still need to understand what is the mechanism behind this lack of electoral consequences, which appears to exist also in consolidated advanced democracies where information is guaranteed and where institutions are relatively strong. In this paper we propose three micro mechanisms that may be behind the limited electoral consequences of corruption. These mechanisms apply to partisans, citizens 3

with a preference for a party, who are significantly more tolerant towards corruption when it affects their own party (Anduiza et al 2014), and thus are expected to be more likely to keep vote for a corrupt politician of their own party. First, voters may perceive that a successful administration, in terms of access to and distribution of resources, compensates the costs of corruption. This explanation was already suggested in the classic work by Rundquist, Strom and Peters (1977) under the label of implicit exchange, and could be reflected in the Brazilian expression rouba mais faz. Second, voters may not give credibility to information about corruption, arguing that it is a baseless accusation of the opposition. This can be labelled as the noise hypothesis, in which political conflict is so pervasive that citizens tend to ignore charges of corruption. Third, voters could consider that all parties are equally affected by corruption and thus, although they may reject corruption practices, corruption may not actually make them change their vote. This can be named the cynicism explanation for reduced electoral consequences of corruption. Our paper provides compelling evidence that the implicit exchange hypothesis is the most important mechanism behind the limited electoral consequences of corruption, and that noise also plays a significant role. With the current design we are unable to assess to what extent cynicism matters. The paper is structured in four sections. The first section discusses previous findings on the limited electoral consequences of corruption and on the arguments why this may be so. Section 2 presents the research design, a survey experiment, with which we test our three different hypotheses. Section 3 presents the results and finally section 4 discusses the findings. Theory The modest electoral consequences of corruption Previous research has found that, contrary to what democratic theory would lead us to expect, corruption modestly reduces incumbents share of the vote and very often results in the candidate being re-elected. Peter and Welch (1980) estimate that US House of 4

Representative candidates charged with corruption loose between 6 and 11% in their expected share of the vote, and, although less likely to be re-elected than uncharged candidates, they are more likely to be re-elected than not. These findings were confirmed with a follow up study that showed a larger effect of corruption on reelection probabilities (Welch and Hibbing, 1987), but still those charged were more likely than not to be re-elected. Dimock and Jacobson (1995) show that although there was a 5% reduction in the incumbents share of the vote when affected by a House of Representatives corruption scandal in 1992, the survival rate was still 80% compared to a 98% of those not affected, an impressive rate in spite of the important number of prior strategic retirements Groseclose and Krehbiel (1994). Eggers & Fisher (2011) assessed the electoral impact of the 2009 UK parliamentary expenses scandal, focusing on whether MPs who were implicated in the scandal retired at a higher rate or received lower electoral support in the 2010 general election than those who were not. They found that implication in the scandal led to a higher retirement rate and a lower vote share for implicated MPs. However the results show that the expenses scandal had a modest impact on constituency-level outcomes. They point to institutional factors as an explanation for this (electoral system and separation of powers). Reed (1999) found that in Japan legislators indicted for corruption only loose a few percentage points and those convicted actually increase the share of the vote, with a large majority being re-elected. Chang, Golden and Hill (2010) examine Italian deputies suspected of criminal wrongdoing and find that only in the 11 th legislature Italians have started to held politicians relatively accountable and remove them from office. Deegan-Krause, Klasnja and Tucker (2011) show with survey evidence, that while personal experience with corruption reduces the probability of voting for an incumbent, but perceived corruption does not. Pocket-book corruption voting is thus much more prevalent than sociotropic corruption voting. However in advanced democracies we do not expect so much personal encounters with corruption practices, but rather, corruption as a contextual factor that may or may not be perceived. The fact that there is no 5

sociotropic corruption is thus worrying. In a recent study Slomczynski and Shabad (2011) have recently found that perceiving a party to be corrupt made voters less likely to vote for that party. However their analysis do not take into account the fact that that individuals that are more likely to vote for a party may perceive lower levels of corruption for this party. The case of Spain seems particularly adequate to analyse the lack of dramatic electoral consequences of corruption. Costas, Solé-Ollé and Sorribas-Navarro in an analysis based on press reports published between 1996 and 2009 show that there are modest electoral punishments for corruption cases in Spanish municipalities (an average of 3%). Some circumstances such as extensive newspaper coverage or direct judiciary intervention can increase these to 9 or 12%. Rivero-Rodriguez and Fernandez-Vazquez (2011) reach the clear and dramatic conclusion that mayors involved in corruption cases in Spain are not punished at the polls, even when taking into account the differences between municipalities affected with corruption and those that are not. Some conditionants: partisanship, information, context These modest effects of corruption on election results has pushed scholars to pay attention to the conditions under which this relationship may be particularly weak (or strong). Partisanship is a crucial variable that has often been considered in the literature. Dimock and Jacobson (1995) show that partisanship moderates the electoral damage of corruption. Voters of the incumbent party are more likely to think than the incumbent is innocent when it is not. When facing the alternative between condemning an incumbent they liked or considering the offense as inconsequential they often chose the latter. Cognitive dissonance may explain these patterns. Davis, Camp and Coleman (2004) show that partisans of opposition parties are more likely to perceive corruption than supporters of the party in power. In a more general perspective, Anderson and Tverdova (2003) show that the negative effect of corruption on evaluations of the political system is attenuated among supporters of the incumbents, and Gonzales et al (1995) shows that partisans provides more favourable evaluations to allegations of political misconduct affecting politicians of the same party. 6

In previous work (Anduiza, Gallego & Muñoz 2014) we have shown that partisans are indeed more likely to be tolerant with the corruption cases that affect their own party. Political sophistication, however, reduces this partisan bias and reduces the tolerance that citizens may have towards corruption practices. Indeed, information is considered as another relevant variable in this picture. If provided with sufficient and credible information about corruption cases, the electoral consequences of corruption would be more clear. Fackler and Lin (1995) show that by including information about corruption together with information about the economy they can better explain presidential elections outcomes. Figueiredo et al (2000) show that information enhances accountability by modesty moving 3% of the vote in the informed vs. the control group. However this happens only in some cases (when a left party is affected) and information also reduces electoral turnout. Cong, De la O, Karlan and Wantchekon (2011) show that information about corruption cases reduces incumbent vote, but also the challenger s votes and turnout in Mexico s local elections. Information alone thus does not guarantee accountability, as voters may respond withdrawing from the political process. But even under the assumption of information voters seem to be relatively reluctant to electorally punish corruption. Finally, the characteristics of the political context may also seem relevant as they affect corruption levels and accountability (Lederman, Loyaza & Soares, 2005) and access to information (Cong et al. 2011). Krause and Mendez (2009) find that increases in perceptions of corruption reduce the incumbent s share of the vote but only in new democracies and parliamentary systems. Lederman, Loyaza & Soares, 2005, and Manzetti & Wilson 2007 also underline the important role of strong institutions. Manzetti and Wilson (2007) argue that people vote for corrupt politicians where clientelism is widespread and institutions are weak, because people are able to secure the delivery of goods by voting. Politicians maintain electoral support by manipulating government institutions to benefit their clientelisc networks. 7

Bagenholm (2009) suggests that corruption has electoral consequences when it is politicised, and thus there is an anti-corruption discourse exploited by political parties. Zegmeister and Zizumbo-Colunga (2012) show compelling evidence that the economic context conditions the punishment that electors inflict in corrupt governments: individuals facing bad collective economic conditions apply a higher penalty to presidential approval for perceived political corruption. So why do voters vote for corrupt politicians? Three causal mechanisms All these previous works try to answer the question of under what conditions is corruption more likely to have the electoral consequences it should have. However, they do not directly address the mechanisms through which people decide to vote for a corrupt candidate, and why this happens even in contexts where information and strong institutions are present. This paper addresses the mechanisms that underlie the fact that people consider voting for a corrupt politicians. These explanations have been suggested in some of the works that have found such limited impact of corruption practices on election outcomes, but to our knowledge have not yet been empirically tested. Implicit exchange Rundquist et al (1977) suggest this mechanism after considering that lack of information or direct incentives as material inducements cannot be the only explanation for corruption voting. Support may depend on the ability to distribute patronage (Manzetti & Wilson 2007), but it may be also be provided by inducements that need not be material, nor direct, nor explicit: citizens may consider other components in their vote choice, and prefer a corrupt candidate with certain other characteristics, positions, or record of past experience as incumbent. These other elements may be perceived to increase citizens well-being and thus counterbalance the potential negative effects of corruption. Rundquist et al. (1977) asses the importance of holding different positions on issues relevant to the voter, but did not test for another element that has been suggested by the literature as relevant: the idea that honest challenger will not deliver the same results in terms of economic development and increases in well-being. This 8

has been suggested by Fernandez-Vazquez and Rivero (2011) as a possible explanation for the startling finding that corruption has no electoral consequences at the local level in Spain. Zechmeister and Zizumbo-Colunga (2012) also argue that citizens have a propensity to trade-off political corruption for economic well being, and thus, when the times are hard, perceptions of corruption bring heavier punishments (see also Rundquist et al 1977:961). Weitz-Shapiro and Winters (2010) assess the consequences of corruption and competence on the probability of voting. Although they argue that information about corruption is more important than the candidates performance (incompetent clean candidates are more likely to be voted than competent corrupt ones), competence doubles the chance that a corrupt candidate will be voted. Our first hypothesis thus states that the propensity to support a corrupt politician will significantly increase if this politician is considered to have a good record of previous political experience that brings economic well-being to its constituency. Noise A second mechanism has to do with how parties and citizens react to charges of corruption. This is related to the credibility of corruption (also suggested by Rundquist et al 1977:957). Even when information about corruption is provided, citizens may discard this information as not credible. This has been acknowledged also by Fernandez-Vazquez and Rivero (2011) and Ferraz and Finan (2008), particularly when partisan attachments are at work. Parties can contribute decisively to generating noise and reducing the credibility of a corruption accusation, therefore affecting their electoral consequences. Our second hypothesis states that the propensity to support a corrupt politician will significantly increase if the party refuses to acknowledge the charges. Cynicism Finally, if all parties or contending candidates are equally perceived as corrupt, there is 9

not reason for a citizen to change its vote. This is suggested by Rundquist et al (1977:956). Fernandez-Vazquez and Rivero 2011 argue in the discussion of their null findings that voters may reject dishonest behavior but still not have reasons to change their mayor provided that they expect that the likelihood of future corruption is at least the same with the opposition in power. Thus, our third hypothesis states that voters are more likely to vote for a corrupt candidate when all parties or alternative candidates are perceived to be also affected by corruption. Data The data are obtained from a survey experiment carried out in Catalunya in April 2012. 1,500 individuals (selected with gender, age and education quotas from a commercial database) were interviewed online. Since the hypotheses regard voters with some degree of partisanship, we first ask participants which party they consider closest to their own ideas. If they answer none, they are further requested to identify a party with which they agree more than with others. In our sample of 1,500 cases 1,102 show some degree of closeness to a party. The individuals that said they were not close to any party did not take the experiment. Participants where then divided into six treatment groups, two for each hypothesis. Participants were shown six different hypothetical vignettes, always referring to the party to which the individual felt closer to. Implicit exchange treatments Successful administration, expected to attenuate the electoral consequences of corruption. [PARTY] mayor, with a highly successful administration, charged with corruption A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. Questioned administration, expected to maximise the electoral consequences of corruption. [PARTY] mayor, with a questioned administration charged with corruption A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. 10

The [PARTY] mayor has been noted for its good management. Under his mandate, the municipality has attracted investment, several long-awaited infrastructure projects have been completed, and taxes have been lowered. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. The mayor [PP / CiU / PSC / ERC / IC-V] stands out for its mismanagement. Under his rule the municipality has not attracted investment, long-awaited infrastructure projects have not been finalized, and taxes have been raised. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. Noise treatments High noise (party denial and accusation), expected to reduce the electoral consequences of corruption. [PARTY] defends one of its mayors charged with corruption and accuses the opposition of lying A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. Yesterday, the regional headquarters of the [PARTY] regretted that the opposition launches baseless attacks against the mayor to gain political advantage and cover their own problems. Therefore, "appropriate measures will be taken" against the representatives of the opposition that have accused the mayor. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. Low noise (party acknowledgement), expected to maximise the electoral consequences of corruption. [PARTY] regrets the behavior of one of its mayors, charged with corruption A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. Yesterday, the regional headquarters of the [PARTY] regretted that mayor has broken the trust the party had placed on him, and said that "appropriate decisions will be taken "once the details of the charges are known. Other party officials have requested the resignation of the mayor and his expulsion from the party. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. Cynicism treatments All parties are accused of corruption, expected to reduce the electoral consequences of corruption. Third mayor of municipality, this time from [PARTY], charged with corruption A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. Only one party is accused of corruption, expected to maximise the electoral consequences of corruption. A [PARTY] mayor charged with corruption A judge has called as defendant in a corruption case the current mayor of the town of XXXXXXXX, who belongs to the [PARTY]. 11

This is the third mayor of this town involved in a corruption scandal. His two predecessors, who belonged to two parties currently in the opposition were involved in cases of illegal party financing, bribery and embezzlement of public funds. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. This is the first corruption scandal affecting the town. His predecessors, belonging to two parties currently at the opposition, have not ever been affected by any corruption scandal. The inquiry focuses on the award without a public tender for the construction of a park to a company that allegedly paid bribes to [PARTY] public officers. The park was awarded for 4.5 million in 2006 and a year later, the budget increased to 9.5 million. The probability of voting is measured with the following question: If the case described above referred to the municipality where you live, what would be the probability that you would vote for this mayor? Respondents could choose a position in a scale from 0 (would not vote in any case) to 10 (would vote for sure). Additionally, we included several manipulation checks for each of the hypotheses. For the exchange hypothesis respondents were asked: Could you assess the mayor s administration during his mandate? They could chose between The administration has been generally positive or The administration has been generally negative. For the noise hypothesis respondents were asked: Could you assess the extent to which the mayor s party recognizes the corruption accusation? Respondents could chose between The mayor's party does NOT recognize the alleged mayor s crime or The mayor's party DOES recognize the alleged mayor s crime. For the cynicism hypothesis respondents where asked: Could you assess whether in this case corruption affects one or several parties? Respondents could chose between Corruption in this municipality affects only the current mayor or Corruption in this municipality affects several political parties. Results Table 1 shows the results of the experiment. These show that two of our experimental manipulations produced the expected outcome. Both the noise mechanism and the 12

implicit exchange produced a substantial and significant increase in the expected probability of supporting the allegedly corrupt mayor in a next election. On the other hand, the cynicism mechanism does not appear to have a clear effect. Table 1: Main results Average probability Difference P-value NB N Noise 2.11 187 1.03*** 0.000 No noise 1.08 173 Exchange 2.32 205 1.46*** 0.000 No exchange 0.86 170 Cynicism 1.06 177 0.05 1.000 No cynicism 1.01 190 NB : One-tailed t-test The average probabilities are in every case very low, with a maximum of 2.3 (on a 0-10 scale) for the exchange condition that underlined the positive record of the mayor and a minimum of 0.86 for the case of a mayor with a poor record. However, even with this skewed distribution, we see important differences across treatment conditions. It is the exchange mechanism that, in our experiment, appears to have the strongest effect producing a change of 0.14 in the probability to support the mayor, while the difference between the two noise conditions stays at 0.10. Figure 1 represents the results. We can say, therefore, that both implicit exchange and noise are mechanisms that explain the electoral performance of allegedly corrupt politicians. Our results are probably subject to some degree of social desirability effect, as people can be expected to reluctantly acknowledge that they would vote for a corrupt politician. Thus, these results are expected to be conservative. According to our estimates, convincing voters that corruption accusations are mere partisan tricks and being able to sell a strong record of accomplishments in office can more than double the loyalty of your partisans if you are an incumbent involved in a corruption scandal. Figure 1: Treatment effects 13

We failed to find the same effect for the cynicism mechanism. Both treatments had the same outcomes. However we cannot discard the cynicism hypothesis, because we could not adequately manipulate the treatment. This can be seen from Table 2, where individuals subject to different treatments showed similar values in the manipulation checks for the cynicism hypothesis (but not for the other two hypothesis). Table 2: Manipulation checks Mayor's party recognizes Mayor's party does not recognize Noise 34 66 No noise 62 38 Good record Poor record Exchange 71 28 No exchange 29 72 More than one party Only one party Cynicism 79 21 No cynicism 70 30 14

Discussion Our paper provides complementary evidence to aggregate analysis of the electoral consequences of corruption, displaying the mechanisms through which voters may support a corrupt politician. We do not assess likelihood of re-election (something that the literature has focused on because of its obvious political implications), but we have shown compelling evidence that both implicit exchange and noise play a crucial role when voters decide not to punish a corrupt politician. A good administration record increases the probability of voting for a corrupt candidate in 14 percentage points. This provides empirical evidence to the intuition that people may to some extent accept to trade off some level of corruption for economic wellbeing. The fact that the party involved does not acknowledge corruption accusations increases the probability of voting in 10 percentage points. This has unsettling implications, as parties seem to get a political benefit from denying accusations of corruption and increasing the levels of political conflict and noise. He have not been able, however, to manipulate perceptions on corruption for different parties in our experimental setting and so we can neither accept nor discard the potential validity of the cynicism mechanism. Further research is required to provide an effective treatment for cynicism, and to asses the extent to which the consequences of implicit exchange or noise for voting may depend on individual characteristics such as political sophistication or ideology. References Anderson, C. J., & Y. V. Tverdova. (2003). Corruption, Political Allegiances, and Attitudes Toward Government in Contemporary Democracies. American Journal of Political Science 47(1), 91-109. Anduiza, E., Gallego, A. & Muñoz, J. (forthcoming in 2014). Turning a blind eye. Experimental evidence of partisan bias in attitudes towards corruption. Comparative Political Studies 15

Bågenholm, A. (2009). Politicizing Corruption The Electoral Impact of anti-corruption discourse in Europe 1983-2007. Quality of Government Working Paper Series 20019:10 Chang, E. C. C., M. A. Golden, & S. J. Hill (2010). Legislative malfeasance and political accountability. Word Politics 62, 2. Chong, A., A. De La O, D. Karlan & L. Wantchekon (2011). Looking Beyond the Incumbent: The Effects of Exposing Corruption on Electoral Outcomes, NBER Working Paper No. 17679 Costas, E., Solé-Ollé, A. & Sorribas-Navarro, P. (2010). Do voters really tolerate corruption? Evidence from Spanish Mayors. Unpublished manuscript: Universitat de Barcelona & Institut d Economia de Barcelona (IEB) Davis, C. L., R. A. Camp, & K. M. Coleman. 2004. The Influence of Party Systems on Citizens' Perceptions of Corruption and Electoral Response in Latin America. Comparative Political Studies 37, 677-703 Della Porta, D., & Vannucci, A. (1999). Corrupt Exchanges: Actors, Resources, and Mechanisms of Political Corruption. New York: Aldine De Gruyter. Dimock, M. A. & G. C. Jacobson (1995). Checks and choices: The House bank scandal s impact on voters in 1992. Journal of Politics 57 (4), 1143 59. Eggers, A. & Fisher, A. C. (2011). Electoral Accountability and the UK Parliamentary Expenses Scandal: Did Voters Punish Corrupt MPs? Working paper. http://dx.doi.org/10.2139/ssrn.1931868 Fackler, T. & T. Lin (1995). Political Corruption and Presidential Elections, 1929 1992. The Journal of Politics 57, 971-993 doi:10.2307/2960398 Ferraz, C. and F. Finan (2008). Exposing corrupt politicians: The effects of Brazil's publicly released audits on electoral outcomes. Quarterly Journal of Economics, 1-44. Groseclose, T. & K. Krehbiel (1994). Golden Parachutes, Rubber Checks, and Strategic Retirements from the 102d House. American Journal of Political Science. 38(1), 75-99 Article Stable URL: http://www.jstor.org/stable/2111336 Jiménez, F. & Caínzos, M. (2004): La repercusión electoral de los escándalos políticos. Alcance y condiciones, Revista Española de Ciencia Política 10, 141-170. Deegan-Krause, K., M. Klasnja, J. Tucker (2011). It's the Bribe, Stupid! Pocketbook vs. 16

Sociotropic Corruption Voting. Paper prepared for the APSA conference. Klasnja, M. (2011). Why Do Malfeasant Politicians Maintain Political Support? Testing the 'Uninformed Voter' Argument. Paper presented at APSA Conference. Available at SSRN: http://ssrn.com/abstract=1901683 Krause, S., & Méndez, F. (2009). Corruption and Elections: An Empirical Study for a Cross Section of Countries.. Economics & Politics, 21(2), 179 200. doi:10.1111/j.1468-0343.2008.00341.x Kurer, O. (2001). Why Do Voters Support Corrupt Politicians? The Political Economy of Corruption. London: Routledge. Lederman, D., Loayza, N.V. and Soares, R.R. (2005): Accountability and corruption: political institutions matter, Economics and Politics 17(1), 1-35. Manzetti, L. & C. J. Wilson. (2007). Why Do Corrupt Governments Maintain Public Support? Comparative Political Studies 40 (8), 949-970. Peters, J. G. & Welch, S. (1980). The Effects of Charges of Corruption on Voting Behavior in Congressional Elections. American Political Science Review, 71(3), 697-708. Reed, S. R. (1999). Punishing corruption: The response of the Japanese electorate to Scandals. Political Psychology in Japan: Behind the Nails Which Sometimes Stick out (and Get Hammered Down). Commack, N.Y.: Nova Science. Rivero-Rodríguez, G., & Fernández-Vázquez, P. (2011), Las consecuencias electorales de los escándalos de corrupción municipal, 2003-2007. Estudios de Progreso. Madrid: Fundación Alternativas. Rundquist, B. S., G. S. Strom & J. G. Peters. (1977). Corrupt Politicians and Their Electoral Support: Some Experimental Observations. The American Political Science Review 71(3), 954-963. Slomczynski, K. M. & G. Shabad. (2011). Perceptions of Political Party Corruption and Voting Behavior in Poland. Party Politics pp. 1-21. doi: 10.1177/1354068810393266 Weitz-Shapiro, R. & Winters, M. (2010). Lacking Information or Condoning Corruption? Voter Attitudes Toward Corruption in Brazil [Online]. SSRN elibrary. Available: http://papers.ssrn.com/sol3/papers.cfm? abstract_id=1641615 17

Welch, S. & J. R. Hibbing (1997). The Effects of Charges of Corruption on Voting Behavior in Congressional Elections, 1982-1990. The Journal of Politics 59 (1): 226-239. Zechmeister, E. and Zizumbo-Colunga, D. (2012) The Varying Political Toll of Corruption in Good versus Bad Economic Times. Paper presented at the LASA conference. 18